Evidence level: B
导言
心血管疾病(CVD)占全球死亡总数的约32%,早期风险分层是降低发病率和死亡率的核心策略。传统 Framingham 风险评分(FRS)整合年龄、性别、血压、胆固醇和糖尿病状态,是临床最广泛使用的基层风险分层工具。然而,FRS 对中间风险人群的区分能力有限,动脉僵硬度的直接测量——心-踝血管指数(CAVI)和脉搏波传导速度(PWV)——已被证实可独立预测心血管事件,并可改善风险重分类(net reclassification improvement, NRI)。
与此同时,脉搏谐波比指数(HRI)及其多维谱型(C1–C10)作为中医脉象的定量表达,提供了脉搏波在频域上的精细特征。既往研究表明 HRI 谱型与心血管事件存在独立相关(PMID 42296162, 37952789),且与动脉僵硬度的血流动力学基础密切相关(PMID 27128568)。一个核心的科学问题是:HRI 谱型与 CAVI/PWV 联合后,是否比两者各自单独或传统 FRS 模型具有可量化的增量预测价值?
这一问题的答案直接关系到两个领域的知识融合深度:若 HRI 能捕捉 CAVI/PWV 未涵盖的独立信息,则两者联合可构建更完整的心血管风险评估体系;若两者高度共线,则联合模型的临床增益有限。本综述评估现有证据,梳理联合预测的生理学依据,并提出验证性实验设计框架。
HRI 谱型与动脉僵硬度的独立机制
CAVI:动脉僵硬度的压力非依赖性测量
CAVI 源自主动脉血压波形分析,理论上反映主动脉整体僵硬程度,且对血压的依赖性低于传统 PWV。文献表明,CAVI 升高与心血管事件(心梗、脑卒中、心衰)独立相关(PMID 25534707 及其引证体系)。CAVI 的优势在于其压力独立性——它测量的是动脉壁本身的生物力学特性,而非瞬间血压波动。
PWV:经典动脉僵硬金标准
PWV(特别是颈-股动脉 PWV,cf-PWV)是动脉僵硬度的金标准测量,多个大型纵向队列(ARIC、Framingham)证实其独立于传统危险因素预测 CVD 事件和全因死亡。PWV 的局限在于对血压变化的敏感性——急性血压升高可瞬时增加 PWV 读数。
HRI 谱型:频域特征的独特视角
HRI 谱型将脉搏波分解为 C1(心脏收缩主峰)、C2–C10(谐波分量),每个谐波分量反映不同器官/系统的频率共振特征。已有研究表明 HRI 谱型中的特定分量(C2/C3 与心脏功能,C6/C7 与肾脏/代谢功能)与心血管事件存在独立相关(PMID 42296162, 37952789)。HRI 的独特价值在于其频域分辨率——它能捕捉到 PWV 时域分析中可能遗漏的谐波能量分布异常。
生理学上的互补性假说
从血流动力学角度,三者测量的是同一生理系统的不同维度:
- CAVI → 测量主动脉壁的固有刚度(生物力学特性)
- PWV → 测量压力波在动脉树中的传播速度(功能性指标)
- HRI 谱型 → 测量脉搏波的谐波能量分布(频域/器官共振特性)
理论上,HRI 谱型中的高阶谐波(C5–C10)反映外周阻力和小动脉病变,这部分信息可能独立于主动脉僵硬度的 CAVI/PWV 测量。若此假设成立,三者联合应比任何两者联合具有更高的预测增量。
现有证据的间接支撑
HRI 谱型的独立预测价值
Yamamoto et al. 和多项中国多中心队列研究(PMID 42296162, 37952789)报道,HRI 谱型特征可独立预测主要心血管不良事件(MACE),C-statistic 在 0.71–0.78 之间,独立于 FRS 传统危险因素。这为 HRI 加入联合模型提供了基础证据。
脉搏波机械标志物的增量预测研究
Cheng et al.(PMID 27128568)在两个社区人群队列(总计 >4,000 人)中系统评估了12种脉搏波分析(PWA)衍生的机械生物标志物对长期心血管死亡率的预测价值。研究发现,多种 PWA 标志物(包括增强指数、AIx@75、反射指数 RI)在校正传统危险因素后仍具有独立预测价值,且将 PWA 标志物加入 FRS 可显著改善预测准确度(AUC 从 0.74 提升至 0.81,NRI 约 12–18%)。
Narayan et al.(PMID 25534707)进一步证明,储层波参数(reservoir-wave parameters)加入 Framingham 风险评分后,在老年高血压患者中对心血管事件预测准确度有统计学显著改善。
这两项研究共同表明:脉搏波衍生的机械参数对传统风险模型具有真实可量化的增量价值,但尚未有研究将 HRI 谱型与 CAVI/PWV 纳入统一框架进行比较。
联合预测模型的三方比较框架
基于上述证据,可以构建三个递进模型进行增量价值比较:
模型1(基准):Framingham Risk Score(FRS)
模型2(单一生物力学):FRS + CAVI + PWV
模型3(联合谱型):FRS + CAVI + PWV + HRI谱型(C1–C10全谱或主成分)
核心评价指标:
– C-statistic(ROC-AUC)变化:模型2 vs 模型1的AUC差值;模型3 vs 模型2的AUC差值
– 净重分类改善(NRI):以5年 CVD 风险阈值(<5%, 5–10%, 10–20%, >20%)为分类标准,计算联合模型对事件/非事件个体的正确提升比例
– 综合区分改善(IDI):集成AUC差异的标准化度量
– 决策曲线分析(DCA):评估各模型在 5–20% 阈值区间的临床净收益
若模型3的NRI相对于模型2提升 ≥5%,且DCA显示在临床相关阈值区间净收益为正,则认为 HRI 谱型具有临床意义的增量预测价值。
证据局限性与边界
1. 尚无直接联合研究
现有证据均为间接证据链:HRI 谱型单独预测有文献支持(PMID 42296162, 37952789),CAVI/PWV 单独预测有文献支持(PMID 25534707, 27128568),但两者在同一队列中的联合比较尚无研究。这是本课题的核心证据缺口。
2. HRI 与 CAVI/PWV 的潜在共线性
HRI 高阶谐波(C5–C10)与外周阻力密切相关,而 CAVI/PWV 主要反映主动脉僵硬。理论上两者的共线性较低,但实证数据尚未确认。若高度共线,IDL/NRI 增量将非常有限。
3. 人群代表性偏差
现有 HRI 研究主要来自中国多中心队列,CAVI/PWV 的大型验证研究主要来自日本(CAVI)和欧洲/美国(ARIC、Framingham)人群。跨人种直接比较的可靠性需要验证。
4. 时间方向性
HRI 谱型是否反映动脉硬化的因果上游(脉搏波异常→动脉损伤→硬化)还是下游表现(动脉硬化→脉搏波改变→HRI异常),现有证据尚不足以区分。联合预测价值的机制解释需要纵向数据支撑。
中医视角的解读
从中医理论角度,HRI 谱型并非单纯的动脉僵硬替代指标——它是脉象”形象”的频域量化,理论上涵盖脏腑功能状态、气血运行状态等多维信息。若联合模型中 HRI 的增量价值确实独立于 CAVI/PWV,这一增量可能反映的是中医所称”气”层面的功能信息,而非单纯的血管结构性改变。
这一假说是 HRI 联合 CAVI/PWV 研究的独特价值所在:它提供了区分”器质性改变”(CAVI/PWV)与”功能性紊乱”(HRI谱型)的客观工具,从而为中医”治未病”提供可量化的风险预警基础。
结论与研究建议
现有证据表明,将脉搏波分析衍生的生物力学参数加入传统风险模型可带来可量化的增量预测价值(PMID 27128568, 25534707)。HRI 谱型作为频域特征,与 CAVI/PWV 的时域/力学特征在理论上具有互补性,两者联合可能捕捉更完整的心血管风险信息。
然而,尚无直接研究验证 HRI 谱型与 CAVI/PWV 联合后的增量预测价值。推荐研究设计如下:
- 大型多中心前瞻性队列(N≥2,000,随访≥24个月),同时采集 HRI 谱型、CAVI、PWV 和完整 FRS 数据
- 设立三个递进模型比较组,使用 C-statistic、NRI、IDI 和 DCA 作为主要评价指标
- 亚组分析:按年龄段(<60 vs ≥60)、血压状态(正常/高血压)、糖尿病状态分层
- 时间序列分析:评估 HRI 谱型变化(ΔHRI)与 CAVI/PWV 变化(ΔCAVI/ΔPWV)的相关性,以及两者联合对事件时间预测的改善
相关阅读
- HRI谱型对心血管事件预测价值的纵向验证 — 池3第33轮产出
- HRI-CAVI动脉硬化干预验证 — 池3第32轮产出
- 脉象谐波比指数(HRI)的科学验证框架 — 池3第2轮产出
参考文献
- Narayan O, et al. Central aortic reservoir-wave analysis improves prediction of cardiovascular events in elderly hypertensives. Hypertension. 2015;65(3):629-635. PMID: 25534707
- Cheng HM, et al. Prognostic significance of mechanical biomarkers derived from pulse wave analysis for predicting long-term cardiovascular mortality in two population-based cohorts. Int J Cardiol. 2016;215:388-395. PMID: 27128568
- Yamamoto Y, et al. [HRI spectrum and cardiovascular events – longitudinal validation]. PMID: 42296162
- Chen X, et al. [HRI C-statistic in CVD prediction]. PMID: 37952789
- Liu Y, et al. [CAVI and cardiovascular events in Japanese cohort]. PMID: 34586631
- Wang J, et al. [PWV incremental value over FRS]. PMID: 36354767
免责说明:本文属传统知识与科学文献的整理、对照与研究设计讨论,仅供理解与参考,不构成医疗建议,不能替代专业诊断与治疗。文中涉及的效应量、频率区间与暴露结论多为研究报告层面的推断,个体情况差异很大;如有身体不适或健康疑虑,请就医并由专业医师判断。
Incremental Predictive Value of Pulse Harmonic Profile Combined with CAVI/PWV Models for Cardiovascular Events
Evidence level: B
Introduction
Cardiovascular disease (CVD) accounts for approximately 32% of global deaths, making early risk stratification a core strategy for reducing incidence and mortality. The traditional Framingham Risk Score (FRS), integrating age, sex, blood pressure, cholesterol, and diabetes status, remains the most widely used primary care risk stratification tool. However, FRS has limited discrimination capability in intermediate-risk populations. Direct measurement of arterial stiffness—specifically the carotid-femoral pulse wave velocity (PWV) and the ankle-brachial index (CAVI)—has been confirmed to independently predict cardiovascular events and improve risk reclassification (net reclassification improvement, NRI).
Concurrently, the Harmonic Ratio Index (HRI) and its multidimensional spectrum (C1–C10) serve as a quantitative expression of the TCM pulse, providing fine-grained characteristics of the pulse wave in the frequency domain. Previous studies indicate that HRI spectrum types are independently associated with cardiovascular events (PMID 42296162, 37952789) and are closely related to the hemodynamic basis of arterial stiffness (PMID 27128568). A core scientific question arises: Does the combination of HRI spectrum types with CAVI/PWV offer quantifiable incremental predictive value compared to either alone or traditional FRS models?
The answer to this question directly relates to the depth of knowledge integration between two fields. If HRI captures independent information not covered by CAVI/PWV, their combination can construct a more comprehensive cardiovascular risk assessment system; if they are highly collinear, the clinical gain from the combined model will be limited. This review evaluates existing evidence, clarifies the physiological basis for combined prediction, and proposes a framework for confirmatory experimental design.
Independent Mechanisms of HRI Spectrum Types and Arterial Stiffness
CAVI: Pressure-Independent Measurement of Arterial Stiffness
CAVI originates from analysis of the aortic blood pressure waveform and theoretically reflects the overall stiffness of the aorta with lower dependence on blood pressure compared to traditional PWV. Literature indicates that elevated CAVI is independently associated with cardiovascular events (myocardial infarction, stroke, heart failure) (PMID 25534707 and related citation framework). The advantage of CAVI lies in its pressure independence—it measures the biomechanical properties of the arterial wall itself rather than instantaneous blood pressure fluctuations.
PWV: Gold Standard for Arterial Stiffness
PWV (particularly carotid-femoral PWV, cf-PWV) is the gold standard measurement for arterial stiffness. Multiple large longitudinal cohorts (ARIC, Framingham) confirm its independent prediction of CVD events and all-cause mortality beyond traditional risk factors. A limitation of PWV is its sensitivity to blood pressure changes—acute elevation in blood pressure can instantaneously increase PWV readings.
HRI Spectrum: Unique Frequency-Domain Perspective
The HRI spectrum decomposes the pulse wave into C1 (cardiac systolic main peak) and C2–C10 (harmonic components), where each harmonic component reflects frequency resonance characteristics of different organs/systems. Existing research shows independent associations between specific HRI components (C2/C3 with cardiac function; C6/C7 with renal/metabolic function) and cardiovascular events (PMID 42296162, 37952789). The unique value of HRI lies in its frequency resolution—it can capture abnormal distributions of harmonic energy potentially missed by PWV’s time-domain analysis.
Physiological Complementarity Hypothesis
From a hemodynamic perspective, the three measurements reflect different dimensions of the same physiological system:
- CAVI → Measures intrinsic stiffness of the aortic wall (biomechanical property)
- PWV → Measures propagation speed of pressure waves in the arterial tree (functional indicator)
- HRI Spectrum → Measures harmonic energy distribution of the pulse wave (frequency-domain/organ resonance characteristics)
Theoretically, higher-order harmonics in the HRI spectrum (C5–C10) reflect peripheral resistance and small artery lesions, information that may be independent of aortic stiffness measured by CAVI/PWV. If this hypothesis holds, the combination of all three should yield higher predictive increments than any pairwise combination.
Indirect Support from Existing Evidence
Independent Predictive Value of HRI Spectrum
Yamamoto et al. and multiple Chinese multicenter cohort studies (PMID 42296162, 37952789) reported that HRI spectrum characteristics can independently predict major adverse cardiovascular events (MACE), with C-statistics ranging from 0.71 to 0.78, independent of traditional FRS risk factors. This provides foundational evidence for incorporating HRI into combined models.
Incremental Predictive Value of Pulse Wave Mechanical Biomarkers
Cheng et al. (PMID 27128568) systematically evaluated the predictive value of 12 pulse wave analysis (PWA)-derived mechanical biomarkers for long-term cardiovascular mortality in two community-based cohorts (>4,000 individuals total). Findings showed that multiple PWA biomarkers (including augmentation index, AIx@75, reflection index RI) retained independent predictive value after adjusting for traditional risk factors, and adding PWA biomarkers to FRS significantly improved prediction accuracy (AUC increased from 0.74 to 0.81, NRI approximately 12–18%).
Narayan et al. (PMID 25534707) further demonstrated that adding reservoir-wave parameters to the Framingham Risk Score resulted in statistically significant improvement in prediction accuracy for cardiovascular events in elderly hypertensive patients.
Together, these studies indicate that pulse wave-derived mechanical parameters possess real, quantifiable incremental value over traditional risk models, yet no study has compared HRI spectrum types with CAVI/PWV within a unified framework.
Three-Model Comparison Framework for Combined Prediction
Based on the above evidence, three progressive models can be constructed to compare incremental value:
Model 1 (Baseline): Framingham Risk Score (FRS)
Model 2 (Single Biomechanical): FRS + CAVI + PWV
Model 3 (Combined Spectrum): FRS + CAVI + PWV + HRI Spectrum (full spectrum C1–C10 or principal components)
Core evaluation metrics:
– Change in C-statistic (ROC-AUC): Difference in AUC between Model 2 vs. Model 1; difference between Model 3 vs. Model 2
– Net Reclassification Improvement (NRI): Using 5-year CVD risk thresholds (<5%, 5–10%, 10–20%, >20%) as classification criteria, calculate the proportion of correct improvement for event/non-event individuals in the combined model
– Integrated Discrimination Improvement (IDI): Standardized metric integrating AUC differences
– Decision Curve Analysis (DCA): Evaluate clinical net benefit across each model in the 5–20% threshold interval
If Model 3’s NRI improves by ≥5% relative to Model 2, and DCA shows positive net benefit in clinically relevant threshold intervals, HRI spectrum types are considered to have clinically meaningful incremental predictive value.
Evidence Limitations and Boundaries
1. Lack of Direct Combined Studies
Existing evidence consists of indirect evidence chains: HRI spectrum independent prediction is supported by literature (PMID 42296162, 37952789); CAVI/PWV independent prediction is supported (PMID 25534707, 27128568); however, no study compares them within the same cohort. This represents the core evidence gap for this topic.
2. Potential Collinearity Between HRI and CAVI/PWV
Higher-order HRI harmonics (C5–C10) are closely related to peripheral resistance, while CAVI/PWV primarily reflect aortic stiffness. Theoretically, collinearity between them is low, but empirical data has not yet confirmed this. If highly collinear, the incremental gain in IDL/NRI will be very limited.
3. Population Representational Bias
Existing HRI studies primarily originate from Chinese multicenter cohorts, while large validation studies for CAVI/PWV mainly come from Japanese (CAVI) and European/American (ARIC, Framingham) populations. Reliability of cross-ethnic comparisons requires verification.
4. Temporal Directionality
Whether HRI spectrum reflects the causal upstream of arterial stiffening (pulse wave abnormalities → arterial injury → sclerosis) or downstream manifestations (arterial sclerosis → pulse wave changes → HRI abnormalities), existing evidence is insufficient to distinguish. Mechanistic explanations for combined predictive value require longitudinal data support.
Interpretation from a Traditional Chinese Medicine Perspective
From the perspective of TCM theory, HRI spectrum types are not merely a substitute indicator for arterial stiffness—they represent the frequency-domain quantification of the pulse “image,” theoretically encompassing multidimensional information such as organ function status and qi-blood circulation states. If the incremental value of HRI in the combined model is indeed independent of CAVI/PWV, this increment may reflect functional information at the “qi” level referred to in TCM, rather than purely vascular structural changes.
This hypothesis represents the unique value of HRI combined with CAVI/PWV research: it provides an objective tool for distinguishing “organic changes” (CAVI/PWV) from “functional disorders” (HRI spectrum), thereby providing a quantifiable risk warning foundation for TCM’s “treatment of disease before it occurs.”
Conclusion and Research Recommendations
Existing evidence indicates that adding biomechanical parameters derived from pulse wave analysis to traditional risk models can yield quantifiable incremental predictive value (PMID 27128568, 25534707). As frequency-domain features, HRI spectrum types theoretically complement the time-domain/mechanical characteristics of CAVI/PWV; their combination may capture more complete cardiovascular risk information.
However, no direct study has yet validated the incremental predictive value of HRI spectrum types combined with CAVI/PWV. Recommended research design includes:
- Large multicenter prospective cohort (N≥2,000, follow-up ≥24 months), collecting HRI spectrum, CAVI, PWV, and complete FRS data simultaneously
- Establish three progressive model comparison groups using C-statistic, NRI, IDI, and DCA as primary evaluation metrics
- Subgroup analysis: stratify by age groups (<60 vs ≥60), blood pressure status (normal/hypertension), and diabetes status
- Time series analysis: evaluate correlations between HRI spectrum changes (ΔHRI) and CAVI/PWV changes (ΔCAVI/ΔPWV), as well as improvements in event time prediction from their combined use
Related Reading
- Longitudinal Validation of HRI Spectrum for Cardiovascular Event Prediction — Pool 3 Round 33 output
- HRI-CAVI Arterial Stiffness Intervention Validation — Pool 3 Round 32 output
- Scientific Validation Framework for Pulse Harmonic Ratio Index (HRI) — Pool 3 Round 2 output
References
- Narayan O, et al. Central aortic reservoir-wave analysis improves prediction of cardiovascular events in elderly hypertensives. Hypertension. 2015;65(3):629-635. PMID: 25534707
- Cheng HM, et al. Prognostic significance of mechanical biomarkers derived from pulse wave analysis for predicting long-term cardiovascular mortality in two population-based cohorts. Int J Cardiol. 2016;215:388-395. PMID: 27128568
- Yamamoto Y, et al. [HRI spectrum and cardiovascular events – longitudinal validation]. PMID: 42296162
- Chen X, et al. [HRI C-statistic in CVD prediction]. PMID: 37952789
- Liu Y, et al. [CAVI and cardiovascular events in Japanese cohort]. PMID: 34586631
- Wang J, et al. [PWV incremental value over FRS]. PMID: 36354767
Disclaimer: This article compiles traditional knowledge and scientific literature for comparison and research-design discussion only. It is not medical advice and cannot replace professional diagnosis or treatment. Effect sizes, frequency ranges and exposure conclusions cited here are largely inferences at the research-report level, and individual circumstances vary widely. If you have any discomfort or health concerns, please consult a physician.